[X] My Assistant

I was referring to comparing your runs [8] vs [15]....haste without infusion vs haste with infusion. The difference is below 2% which is not how it is supposed to be (expected ~6% or more). It suggests to me that your run [15] had abnormally bad luck, so that is the first one I would re-test. A complete re-test of runs [5], [7], [8], and [15] would be better for fairness.

When I tested these same things earlier (at my lower level) on numerous occasions I had a run that was significantly faster without infusion (everything else same). So I experienced the same problem as you, and I either needed to re-test or partially ignore those results. In my case, since I tested multiple arenas, I just chose to partially ignore those results and rely on the other data a bit more.

DwD isn't like regular arenas, especially regarding dark infusion vs elec infusion (even on Saturday) because of School Girls being very weak to dark. So I think it's not safe to assume that performance extends to arena 80. Though I don't know for sure how dark and electric compares on arena 80. I only estimated that the difference could be up to 2% turns (not elec vs dark damage) but it might also be less or zero.

As you suggested, I repeated some tests and added a few more. (IMG:[invalid] style_emoticons/default/wink.gif) In fact, I was already forgetting to post the results (they had been completed for some time). (IMG:[invalid] style_emoticons/default/heh.gif)

However, I am already saying that the conclusions I made at that time did not change at all. (IMG:[invalid] style_emoticons/default/sleep.gif)

Haste and Neutral infusion (Dark)
(IMG:[imgur.com] https://imgur.com/zq4GuUj.jpg)

Haste without infusion
(IMG:[imgur.com] https://imgur.com/Iou2gln.jpg)

No Haste and Infusion of the day
(IMG:[imgur.com] https://imgur.com/yVPaHLG.jpg)

No Haste and Neutral infusion (Dark)
(IMG:[imgur.com] https://imgur.com/X9ZYc9W.jpg)

No Haste and No Infusion
(IMG:[imgur.com] https://imgur.com/UCXvCmv.jpg)

And as you can see from the results of the third and fourth tables, there is practically no difference in the type of infusion that is used, regardless of the day of the week. Really, Dark and Holy Infusions only make a difference in SG arenas.

Thanks, this is amazing data! I will only use your [average] column for the following discussion, but it's an average of 5 trials, so now the same number of samples as my own haste tests last year.

And as you can see from the results of the third and fourth tables, there is practically no difference in the type of infusion that is used, regardless of the day of the week. Really, Dark and Holy Infusions only make a difference in SG arenas.

Your result is actually a 1625/1611 ≈ 1% difference which matches what I predicted earlier from theoretical considerations. I had actually stated "expected to be 2% or less" and I indeed expected less; I hadn't showed the math anyway. It's a tiny difference but should be excluded from haste tests because it's large enough to skew the already-close results.

However, I am already saying that the conclusions I made at that time did not change at all.

Thus excluding that group of elec infusion data, let's study your main data.

Your new data shows that with neutral dark infusion, haste worsens your turns by 1690/1625 = 4% exactly, this is down from your initial result of 5.5% worsening, as expected. I suspect your true value might be smaller still. In comparison my haste tests at lower level had 1%~3%.

Your new data shows that without infusion, haste worsens your turns by 1734/1688 = 2.7% which is up from your initial result of 1%. I don't know what your true value should be. In comparison my haste tests at lower level had 4%~6%.

In comparison our datas are reversed from each other, which means we both need more trials to get the true answer. (I am not suggesting we do more trials. It's too much work, I think we have done enough. And my level is not the same as before.)

In theory, haste should perform better with infusion, because infusion makes your main hand do more damage and counters are less important. This matches my data, while Basara is backwards. However, my data might have too much difference. I can't think of any game reason why the true result would be Basara's way.

In theory, haste might perform slightly better at higher level, compared to on a low level player. Because high level players have more enemy PMit to deal with, and also hit the counter attack cap more often. But I'm not absolutely sure, because high level players also have slightly more counters in total.

So far our data suggests that Basara's haste is slightly better than mine, but it's hard to tell. Haste is still bad enough that a high level player shouldn't use it, and it should be even worse on most (smaller swarms) arenas and (larger swarms) grindfest, according to my tests and theory.

Is there anyone who played holy mage with charged set in non-imperil PFUDOR Grindfest? It seems these works better than imp element mage mode, at least when someone have good profs, isn't it?

My friends test it recently, with only PFOH and charged holy set. And the result is amazing, pertty good, completely over my charged snowman set (IMG:[invalid] style_emoticons/default/sad.gif) We have the same DD level, both full forged, the only difference is that his basic prof is much higher, about 30 over mine.

emmm, I just find 23th is the day of holy, which gives him additional holy damage by reduce monster's 10% holy resist.

This post has been edited by Ming28561: Aug 24 2020, 05:15

It's been about 4 and a half months since the last monster analysis, so here's some new data.

This time, I expanded my recording tool to gather more information from the combat log, so there's some additional data I'll put in the next post. You should see that post if you want some more information on the damage-related statistics in this post.

For now, just the standard follow-up from the last time:

CODE

24/10/19 - 2,513 unique monsters, 167,318 total appearances, 1712 avg plvl (sssss2 data)
23/12/19 - 2,631 unique monsters, 175,606 total appearances, 1739 avg plvl (my data, 1 too many runs)
08/02/20 - 2,672 unique monsters, 167,241 total appearances, 1753 avg plvl (sssss2 data)
12/04/20 - 2,686 unique monsters, 167,225 total appearances, 1781 avg plvl (my data)
29/08/20 - 2,790 unique monsters, 167,268 total appearances, 1821 avg plvl (this post)

The total number of unique monsters seen was 2790.
I fought a total of 167,268 monsters. ('Total appearances'.)
The average power level per appearance was 1821. (Total power level / appearances.)
The average power level per unique monster was 1612. (Total power level / unique monsters.)
178 different players have monsters high level enough to show in max-level PFests.
The number of appearances made by level 2250 monsters was 30,948. This is 18.5% of all appearances.
Highly subjective stat (see next post): the average 'damage score' per appearance was 740.

Player ranking data:

CODE

|   Rank | Trainer          |   Monsters |   Counted |   % of Total Apps |   Avg/Count |   Avg/Monster |   Total pLvl |   Total Dmg |   % of Total Dmg |   Dmg/Count |
|--------|------------------|------------|-----------|-------------------|-------------|---------------|--------------|-------------|------------------|-------------|
|      1 | Sushilicious     |        200 |     16561 |             9.901 |        1728 |          1720 |       344036 |     8598430 |            6.943 |         519 |
|      2 | FreeloaderV      |        190 |     14836 |             8.87  |        1784 |          1727 |       328158 |    11344306 |            9.16  |         764 |
|      3 | morineko         |        200 |     14641 |             8.753 |        1741 |          1673 |       334659 |    12879751 |           10.4   |         879 |
|      4 | Nero-Arc         |        200 |     14563 |             8.706 |        1692 |          1687 |       337523 |    10679902 |            8.624 |         733 |
|      5 | serorin          |        105 |     10564 |             6.316 |        1697 |          1694 |       177927 |     8909586 |            7.194 |         843 |
|      6 | sssss2           |        200 |     10177 |             6.084 |        2074 |          1600 |       320040 |     9728638 |            7.855 |         955 |
|      7 | gc00018          |         96 |      7892 |             4.718 |        1736 |          1667 |       160102 |     7346038 |            5.932 |         930 |
|      8 | Petal_Kiss       |        200 |      6058 |             3.622 |        2000 |          1552 |       310443 |     4032490 |            3.256 |         665 |
|      9 | MidNightPass     |        100 |      5870 |             3.509 |        1655 |          1611 |       161142 |     3980743 |            3.214 |         678 |
|     10 | NerfThis         |        141 |      4534 |             2.711 |        2006 |          1506 |       212427 |     3949960 |            3.189 |         871 |
|     11 | lestion          |         43 |      4338 |             2.593 |        1822 |          1675 |        72057 |     3506737 |            2.832 |         808 |
|     12 | moanim           |         60 |      3332 |             1.992 |        1617 |          1585 |        95118 |     2161449 |            1.745 |         648 |
|     13 | Koaen            |          7 |      2997 |             1.792 |        2214 |          2209 |        15469 |     2717166 |            2.194 |         906 |
|     14 | danixxx          |         15 |      1992 |             1.191 |        1723 |          1688 |        25331 |     1434426 |            1.158 |         720 |
|     15 | ????             |          9 |      1810 |             1.082 |        2073 |          1827 |        16445 |      910176 |            0.735 |         502 |
|     16 | .@_@.@_@.        |         23 |      1645 |             0.983 |        1787 |          1578 |        36296 |     1155932 |            0.933 |         702 |
|     17 | RoadShoe         |        150 |      1570 |             0.939 |        1728 |          1480 |       222121 |      419592 |            0.339 |         267 |
|     18 | kamio11          |         10 |      1563 |             0.934 |        1878 |          1696 |        16962 |      546189 |            0.441 |         349 |
|     19 | DJNoni           |         35 |      1557 |             0.931 |        1565 |          1565 |        54787 |      520928 |            0.421 |         334 |
|     20 | Honeycat         |          4 |      1412 |             0.844 |        2233 |          2055 |         8220 |      661992 |            0.535 |         468 |
|     21 | Mantra64         |         14 |      1390 |             0.831 |        1759 |          1645 |        23033 |      549133 |            0.443 |         395 |
|     22 | kzh125           |          5 |      1367 |             0.817 |        2203 |          1944 |         9720 |     1596404 |            1.289 |        1167 |
|     23 | threekoala       |         24 |      1303 |             0.779 |        1932 |          1582 |        37978 |      356530 |            0.288 |         273 |
|     24 | StonyCat         |          3 |      1264 |             0.756 |        2176 |          2161 |         6483 |      901539 |            0.728 |         713 |
|     25 | hellweekdays     |         11 |      1243 |             0.743 |        1670 |          1668 |        18354 |      387456 |            0.313 |         311 |
|     26 | zhenterzzf       |          5 |      1165 |             0.696 |        2052 |          1821 |         9109 |     1276744 |            1.031 |        1095 |
|     27 | jantch           |          9 |      1061 |             0.634 |        1686 |          1679 |        15114 |      300304 |            0.242 |         283 |
|     28 | qdjseh001        |         90 |       996 |             0.595 |        1502 |          1502 |       135190 |      469652 |            0.379 |         471 |
|     29 | ddwiki           |          4 |       992 |             0.593 |        2006 |          1876 |         7504 |     1177225 |            0.951 |        1186 |
|     30 | 15112006         |         14 |       982 |             0.587 |        1626 |          1604 |        22456 |      213291 |            0.172 |         217 |
|     31 | another planet   |          2 |       973 |             0.582 |        2250 |          2250 |         4500 |     1019751 |            0.823 |        1048 |
|     32 | Void Domain      |          2 |       949 |             0.567 |        2250 |          2250 |         4500 |      894293 |            0.722 |         942 |
|     33 | warachiasion     |          3 |       784 |             0.469 |        1961 |          1899 |         5697 |      900313 |            0.727 |        1148 |
|     34 | EsotericSatire   |          3 |       779 |             0.466 |        2040 |          1898 |         5696 |      287922 |            0.232 |         369 |
|     35 | Ichy             |          3 |       752 |             0.45  |        2063 |          1877 |         5632 |      648572 |            0.524 |         862 |
|     36 | Godde??          |         69 |       743 |             0.444 |        1512 |          1506 |       103932 |      164498 |            0.133 |         221 |
|     37 | foobarwtf        |          2 |       671 |             0.401 |        2094 |          2016 |         4033 |      651935 |            0.526 |         971 |
|     38 | KINOSHITAMIKOTO  |          5 |       669 |             0.4   |        2048 |          1731 |         8658 |      504538 |            0.407 |         754 |
|     39 | fihero           |          2 |       654 |             0.391 |        2009 |          1840 |         3681 |      362189 |            0.292 |         553 |
|     40 | Hina Amano       |        110 |       643 |             0.384 |        1507 |          1469 |       161619 |      307103 |            0.248 |         477 |
|     41 | mouisaac         |         16 |       591 |             0.353 |        1560 |          1560 |        24960 |      239089 |            0.193 |         404 |
|     42 | ashimoto         |          3 |       557 |             0.333 |        1778 |          1766 |         5298 |      120004 |            0.097 |         215 |
|     43 | HTTP/308         |          5 |       506 |             0.303 |        2191 |          1648 |         8240 |      537659 |            0.434 |        1062 |
|     44 | Dreamophobia     |          1 |       496 |             0.297 |        2250 |          2250 |         2250 |      527564 |            0.426 |        1063 |
|     45 | arialinnoc       |          6 |       490 |             0.293 |        2053 |          1601 |         9606 |      507838 |            0.41  |        1036 |
|     46 | SakiRaFubuKi     |          1 |       480 |             0.287 |        2250 |          2250 |         2250 |      644974 |            0.521 |        1343 |
|     47 | tibus1979        |          1 |       476 |             0.285 |        2250 |          2250 |         2250 |      220764 |            0.178 |         463 |
|     48 | Ming28561        |          1 |       463 |             0.277 |        2250 |          2250 |         2250 |      440212 |            0.355 |         950 |
|     49 | biborobi         |          1 |       462 |             0.276 |        2250 |          2250 |         2250 |      495789 |            0.4   |        1073 |
|     50 | eliudnir         |          1 |       462 |             0.276 |        2250 |          2250 |         2250 |      334330 |            0.27  |         723 |

This post has been edited by lestion: Oct 13 2020, 04:24

Follow-up to the previous post: I mentioned that I'd tracked some more stuff in the combat log, and that included the damage dealt to me by monsters, the damage I dealt to them, and how long they survived (in time units!).

The 'damage score' I mentioned previously is the damage they actually dealt, normalized to account for the grindfest round damage increase.

Of course, this is not a -great- measure of how dangerous monsters are in general - because it's a measure of what monsters could deal to me, as a highly-geared imperil elec mage. This matters both because of armor type and because mages kill fast enough to avoid SP attacks 99% of the time, and also the distribution of MP attacks vs normal attacks is likely lower. So take this data with a pinch of salt: it's a representation of what monsters did against a high-level imperil mage, not what they can do against all players.

Now I can give some more stats - a breakdown of monster classes. I'm only going to do this post once (partly because of the problem I mentioned above - it's not objective data, and I don't want it to mislead people into making decisions based on it).

BTW, you might need to make your browser window wider if you can't read this table properly, or zoom out... I apologise for the ridiculous width.

CODE

| Class     |   Count |   PLVL/Count |   Unique |   PLVL/Mon |   Avg SurvTime |   Avg Underkill |   Avg Overkill |   Avg dHits |   Avg dDmg |   Avg rHits |   Avg rDmg |   Total Sparks |   Avg SparkRound |
|-----------|---------|--------------|----------|------------|----------------|-----------------|----------------|-------------|------------|-------------|------------|----------------|------------------|
| Giant     |   24045 |         1890 |      356 |       1657 |        231.666 |           20015 |         126239 |       1.652 |        845 |       1.386 |     131420 |            176 |              610 |
| Celestial |   18910 |         1859 |      282 |       1628 |        219.147 |            7970 |         127070 |       1.472 |        529 |       1.212 |     130024 |             63 |              741 |
| Dragonkin |   18041 |         1830 |      303 |       1631 |        233.487 |           18411 |         119233 |       1.682 |        963 |       1.39  |     124004 |            120 |              667 |
| Undead    |   16721 |         1821 |      270 |       1621 |        235.519 |           22094 |         117306 |       1.748 |        981 |       1.435 |     124634 |            130 |              672 |
| Arthropod |   16549 |         1777 |      328 |       1603 |        233.642 |           22299 |         117559 |       1.709 |        814 |       1.427 |     121832 |            103 |              696 |
| Mechanoid |   12304 |         1814 |      236 |       1586 |        221.153 |            7742 |         141805 |       1.491 |        847 |       1.189 |     145100 |             67 |              716 |
| Elemental |   10325 |         1804 |      171 |       1602 |        212.511 |            5822 |         133578 |       1.328 |        383 |       1.154 |     133776 |             34 |              821 |
| Humanoid  |    9520 |         1876 |      120 |       1628 |        219.451 |            7456 |         126933 |       1.476 |        578 |       1.196 |     134803 |             27 |              768 |
| Beast     |    9399 |         1811 |      154 |       1604 |        223.94  |           10060 |         127349 |       1.552 |        761 |       1.251 |     131844 |             46 |              758 |
| Sprite    |    8948 |         1821 |      146 |       1590 |        213.505 |            4447 |         142400 |       1.346 |        468 |       1.123 |     135704 |             21 |              780 |
| Daimon    |    8659 |         1772 |      152 |       1585 |        219.201 |            6031 |         137186 |       1.467 |        799 |       1.159 |     132128 |             52 |              736 |
| Reptilian |    7551 |         1705 |      138 |       1589 |        223.525 |           11688 |         124518 |       1.54  |        656 |       1.294 |     129169 |             20 |              770 |
| Avion     |    6296 |         1723 |      134 |       1566 |        214.099 |            4746 |         138878 |       1.358 |        503 |       1.129 |     136072 |             18 |              728 |

This data is sorted by dDmg, because I was trying to model which monsters are the most dangerous to me.

Explanations:
'SurvTime' is time units survived. Higher is better. Don't worry about how the exact numbers are gathered if you aren't familiar with the action speed calculation - it's not important.
'Underkill' is how much health each monster had left on the turn before I killed it.
'Overkill' is as it sounds - how much more damage I did to it than I needed to reduce it to exactly 0.
'dHits' and 'dDmg' - average times the monster hit me per round, and how much damage they did with each hit.
'rHits' and 'rDmg' - same as above except it's my hits against them.
Sparks - how many times they triggered my spark of life.
SparkRound - the rounds at which they triggered it. (Lower is better here - means they started being dangerous earlier in a grindfest.)

Also, here's the data for 2250s only:

CODE

| Class     |   Count |   PLVL/Count |   Unique |   PLVL/Mon |   Avg SurvTime |   Avg Underkill |   Avg Overkill |   Avg dHits |   Avg dDmg |   Avg rHits |   Avg rDmg |   Total Sparks |   Avg SparkRound |
|-----------|---------|--------------|----------|------------|----------------|-----------------|----------------|-------------|------------|-------------|------------|----------------|------------------|
| Giant     |    5723 |         2250 |       14 |       2250 |        239.365 |           29560 |         113463 |       1.795 |        998 |       1.516 |     121311 |             46 |              689 |
| Celestial |    3830 |         2250 |        9 |       2250 |        223.058 |           10178 |         117593 |       1.582 |        675 |       1.279 |     127933 |             18 |              801 |
| Dragonkin |    3393 |         2250 |        8 |       2250 |        240.964 |           27902 |         109085 |       1.885 |       1209 |       1.524 |     120599 |             29 |              749 |
| Undead    |    3144 |         2250 |        7 |       2250 |        240.5   |           28183 |         110440 |       1.844 |       1136 |       1.491 |     125728 |             30 |              735 |
| Mechanoid |    2989 |         2250 |        7 |       2250 |        226.57  |           12269 |         130257 |       1.679 |       1092 |       1.294 |     141185 |             18 |              781 |
| Humanoid  |    2879 |         2250 |        6 |       2250 |        225.036 |           10307 |         121258 |       1.544 |        709 |       1.261 |     132868 |             13 |              735 |
| Sprite    |    2297 |         2250 |        5 |       2250 |        219.957 |            6876 |         122234 |       1.571 |        651 |       1.192 |     130847 |              6 |              731 |
| Arthropod |    1850 |         2250 |        4 |       2250 |        241.173 |           33558 |         106723 |       1.943 |       1030 |       1.569 |     118605 |             14 |              681 |
| Beast     |    1813 |         2250 |        4 |       2250 |        231.85  |           15453 |         121419 |       1.755 |        944 |       1.365 |     126703 |              8 |              744 |
| Elemental |    1690 |         2250 |        4 |       2250 |        218.133 |            9649 |         118495 |       1.507 |        548 |       1.244 |     129099 |              6 |              911 |
| Daimon    |    1340 |         2250 |        3 |       2250 |        223.192 |            9872 |         117204 |       1.642 |       1052 |       1.243 |     128688 |             11 |              757 |

Finally I have a little more data to give in a final post...

This post has been edited by lestion: Aug 30 2020, 13:07

Finally...

I can also analyse an individual player's monsters. If you are in the top 50 and you want your monster data, feel free to PM me. It looks like this:

CODE

|   Rank | Name                         |   PL | Class     |   Count |   SurvTime |   Dmg/Count |   Sparks |
|--------|------------------------------|------|-----------|---------|------------|-------------|----------|
|      1 | Fist Of The Porn Star        | 2250 | Sprite    |     496 |     220.6  |         698 |        1 |
|      2 | Hardcore Levelling Warrior   | 2250 | Giant     |     446 |     250.14 |        1107 |        6 |
|      3 | Thomas The Spank Engine      | 1809 | Mechanoid |     227 |     232.1  |        1017 |        1 |
|      4 | 10boros Bazaar Adventure     | 1728 | Daimon    |     161 |     226.67 |         961 |        1 |
|      5 | Hijikata Toshiro             | 1728 | Daimon    |     149 |     224.39 |         714 |        0 |
|      6 | Cum Captor Sakura            | 1728 | Elemental |     143 |     214.9  |         435 |        0 |
|      7 | Doki Doki Erotica Club       | 1728 | Arthropod |     143 |     244.94 |         968 |        1 |
|      8 | Bukkakemonogatari            | 1728 | Undead    |     142 |     241.29 |        1181 |        1 |
|      9 | Kondo Isao                   | 1728 | Dragonkin |     141 |     245.93 |        1102 |        0 |
|     10 | Jesus Christ                 | 1728 | Undead    |     140 |     248.08 |        1157 |        1 |
|     11 | Sanya V Litvyak              | 1728 | Celestial |     140 |     216.51 |         453 |        1 |
|     12 | Miyafuji Yoshika             | 1728 | Celestial |     140 |     222.82 |         680 |        1 |
|     13 | Charlotte E Yeager           | 1728 | Celestial |     134 |     219.02 |         501 |        0 |
|     14 | Sakamoto Mio                 | 1728 | Celestial |     132 |     211.28 |         501 |        2 |
|     15 | Nikka Edvardine Katajainen   | 1728 | Celestial |     124 |     218.05 |         582 |        0 |
|     16 | Sakata Gintoki               | 1728 | Beast     |     124 |     221.81 |         797 |        0 |
|     17 | Gertrud Barkhorn             | 1728 | Celestial |     120 |     227.9  |         691 |        0 |
|     18 | Lynette Bishop               | 1728 | Celestial |     117 |     221.08 |         534 |        0 |
|     19 | Eila Ilmatar Juutilainen     | 1728 | Celestial |     117 |     229.07 |         551 |        0 |
|     20 | Minna-dietlinde Wilcke       | 1728 | Celestial |     117 |     218.26 |         662 |        2 |
|     21 | Francesca Lucchini           | 1728 | Celestial |     116 |     226.7  |         623 |        0 |
|     22 | Erica Hartmann               | 1728 | Celestial |     114 |     217.55 |         553 |        1 |
|     23 | Fullmetal Alcummies          | 1638 | Mechanoid |      89 |     213.81 |        1289 |        0 |
|     24 | K-onahole                    | 1638 | Giant     |      58 |     235.35 |        1027 |        0 |
|     25 | Satan Himself                | 1560 | Daimon    |      47 |     203.68 |         333 |        0 |
|     26 | A Certain Pornographic Index | 1560 | Elemental |      39 |     210.84 |         296 |        0 |
|     27 | Little Bitch Academia        | 1560 | Daimon    |      37 |     212.12 |         882 |        0 |
|     28 | Fill Me Baby                 | 1560 | Sprite    |      36 |     208.2  |         665 |        0 |
|     29 | Median Joe                   | 1560 | Undead    |      32 |     250.74 |        1180 |        2 |
|     30 | Judas Iscariot               | 1560 | Daimon    |      31 |     213.94 |         707 |        0 |
|     31 | Extreme Joe                  | 1560 | Undead    |      29 |     235.52 |         879 |        0 |
|     32 | Speeling Errer               | 1560 | Arthropod |      29 |     227    |         922 |        0 |
|     33 | Fellator Of Snowflake        | 1560 | Giant     |      28 |     230.49 |        1400 |        0 |
|     34 | Largest Joe                  | 1560 | Undead    |      27 |     237.38 |         946 |        0 |
|     35 | Utmost Joe                   | 1560 | Undead    |      24 |     232    |        1325 |        0 |
|     36 | Kim Jong Un-dead             | 1560 | Undead    |      22 |     248.41 |        1090 |        0 |
|     37 | Lucky Shart                  | 1560 | Giant     |      21 |     249.38 |        1392 |        0 |
|     38 | Covid-19                     | 1560 | Celestial |      21 |     214.04 |         299 |        0 |
|     39 | Homo Erectus                 | 1560 | Giant     |      20 |     217.42 |        1040 |        0 |
|     40 | One Penis Man                | 1560 | Giant     |      19 |     239.1  |         801 |        0 |
|     41 | Standard Deviation Joe       | 1560 | Undead    |      19 |     278.27 |        1196 |        0 |
|     42 | Sailor Poon                  | 1560 | Celestial |      15 |     255.33 |         659 |        0 |
|     43 | Read The Fucking Wiki        | 1560 | Celestial |      12 |     228.83 |         553 |        0 |

awesome work, lestion!

research

You are useful and smart (IMG:[invalid] style_emoticons/default/wub.gif)

-snip-

(IMG:[i.imgur.com] https://i.imgur.com/9Z9342a.jpg)

This will be 10MC. (IMG:[invalid] style_emoticons/default/happy.gif)

This post has been edited by decondelite: Sep 1 2020, 20:22

I bow to all you, for doing such in depth research. Thank you, everyone of you, for your dedication and generosity sharing it all.

There is a wealth of knowledge embedded here, but I am not certain I can follow it all, even reading iit all (and I've only skimmed the last few pages.

I don't suppose that there is a summary or list of important implications? If not, could I offer someone some credits to write one up? I'm sure it's not sufficient to truly reward any of you for the effort and thought that has gone into this, but I would offer 100K for a clear and detailed write up of the implications, particularly for 1H style power.

Or is that insultingly low? I hope not.

I don't need to understand the research, so much as what the implications are. Maybe something that was structured into a list of short topics?

• The benefits ascribed to this common advice (x) is actually not supported by the research. The implication is this...
• The research tells us that x,y,z while counter-intuitive, is actually true. The implications are these...
• Haste changes the outcomes a,b,c ways in x,y,z circumstances/play styles/battle modes. The trade offs are these...
• Imperil and infusions have these effect, in these circumstances, for this percentage additional cost.
• Modify your play style these ways, on these days of the week, for this percentage extra cost/time/whatever, results in these increases/decreases in x,y,x.
• These are the most common things for which players optimize (speed of leveling, increased income/lowered cost, ease of play, other categories I haven't thought of, but that you may consider foundational knowledge) and the implications of all this in depth research for each is...
• Optimizing for a,b,c, follow these best practices, but these are the trade offs...
• The research tells us that the highest returns (on a, b, c) are gained by these optimizations, while these others have measurable, but lower benefits, and are only worth while under these circumstances.
• Never do x, except for these exceptions.

Or whatever the implications are, to be clear.

I respect you all, tremendously, and hope that asking like this (and my paltry 100K offer) does not offend. If 100K isn't a worthwhile offer for the effort to summarize, I completely understand.

Perhaps what I really want is to pay one of you dedicated researchers/experts, who understand all the details and implications, for a consultation, helping me optimize my play style.

But the offer stands if someone were willing/able to boil all this down into just the important implications. I wish I were rich, and could reward all the wonderful research.

Thank you!

I decided to take another stab at coming up with the formula to calculate PXP0 and the bazaar price of equipments using their individual stats. This has been tried before by various players and more clues are revealed each time.

Can people who own any of these weapons please check their bazaar sell prices?

Peerless Ethereal Rapier of Slaughter
Peerless Non-Ethereal Rapier of Slaughter
Peerless Ethereal Shortsword of Slaughter
Peerless Non-Ethereal Shortsword of Slaughter

All of those should have 377 PXP0. I want to know the bazaar sell prices for all four cases.

I would also be interested to know the bazaar sell prices for any of those with the Balance suffix instead. Note all of those should have 378 PXP0.

Minor test. I equipped a 1H wakizashi with Overpower 5 (+20% counter-parry) and tried playing through the first six arenas carefully, checking turn-by-turn if I ever got parried with 3~4 stacks of Overwhelming Strikes. I did get parried once with 3 stacks, but never with 4 stacks.

I played in such a way as to maximize the time I have 4 stacks (deathblow enemies to prevent reaching 5 stacks, and resetting with defend as soon as I did) and not hitting stunned enemies with 4 stacks. Although not absolutely definitive with this number of trials, the result suggests counter-parry stacks additively in all instances. So Overpower 5 plus Overwhelming Strikes 4 is probably indeed 100% counter-parry (as I have been assuming in the past).

Someone who writes custom scripts could probably verify this in grindfest or something.

Has anyone investigated the distribution of low/mid/high-grade materials?
I tried to calculate the distribution based on the assumptions:
1 Chance of Low-Grade is 100% at zero level, and linearly falls to 0% at 750 level.
2 Chance of High-Grade is 0% at level 500, and increases linearly to level 1500.
3 Chance of Mid-Grade = 100% - Chance of Low-Grade - Chance of High-Grade.

I have 76 monsters in the lab mostly 531 levels, according to calculations distribution should be approximately 29,2/67,7/3,1 Low/Mid/High.
According to my log it comes out 24,9/71,6/3,5
It seems to be about the same, even taking into account that I have several monsters above level 531.

Can I upload Excel files here? Uploaded to mega.
[mega.nz] https://mega.nz/file/EzAnlSaD#iAaRzh_Nco5JA...varsrQjhLm11cWw

I decided to take another stab at coming up with the formula to calculate PXP0 and the bazaar price of equipments using their individual stats. This has been tried before by various players and more clues are revealed each time.

Can people who own any of these weapons please check their bazaar sell prices?

Peerless Ethereal Rapier of Slaughter
Peerless Non-Ethereal Rapier of Slaughter
Peerless Ethereal Shortsword of Slaughter
Peerless Non-Ethereal Shortsword of Slaughter

All of those should have 377 PXP0. I want to know the bazaar sell prices for all four cases.

I would also be interested to know the bazaar sell prices for any of those with the Balance suffix instead. Note all of those should have 378 PXP0.

The bazaar sell price for the exactly same item varies also with his condition. To be comparables, you should specify that you want the bazaar price of a fully repaired item (i.e., condition 100%).

For instance, my Peerless Arctic Shortsword of Slaughter has a bazaar sell price of 4161 c @ 77% condition.

It would have a bit more if I repaired it, and a bit less if I let it go down to 50%.

This post has been edited by mundomuñeca: Nov 17 2020, 14:34

I knew about that, just didn't mention it. I think the bazaar price simply goes up linearly with the current condition (don't think if it's linear with maximum condition though).

How much exact condition does your shortsword have right now? Is it 290/377 or 291/377?

My guess is that after you repair it, the bazaar sell price will be around 5404 credits (don't know which way rounding will go) but we'll see I hope. :)

By the way, I forgot that soulfused/lottery equipment can't be sold to the bazaar (not sure) so I guess we can't check the sell prices using those.

This post has been edited by BlueWaterSplash: Nov 18 2020, 07:38

I should have posted these results before, but the laziness was stronger. (IMG:[invalid] style_emoticons/default/heh.gif)

As I did a lot of tests, I will divide it into three parts.

Part 1: Comparison between different "1H heavy sets".

(IMG:[imgur.com] https://imgur.com/XgDwtAZ.jpg)

Arena chosen: 150 rounds
Equipment forge: Rapier Slaughter, Shield and Slaughter Set (full forge); Balance, Protection, Warding and Shielding Sets (forge level 25)
Hath Perk: Dæmon Duality V (no other "game perk" more)
Supportive Spells in continuous use: Regen, Protection, Spirit Shield, Spark of Life, Heartseeker
Deprecating Spell in continuous use: Imperil
Shock Spike Shield
Use of Voidseeker Shard in the Shielding set, to increase the hit chance above 200%.

(IMG:[imgur.com] https://imgur.com/mvBNVtM.jpg)

Although the statistics of the sets are not exactly from the same time I performed the tests, the numbers have not changed significantly.

The test dates, for the same set, are not continuous because in some cases I had to retake the test due to connection problems or inconsistent results (very different from the others).

Links of my equipment can be found in my signature.

[ 1 ] Balance Set

(IMG:[imgur.com] https://imgur.com/Ig7YOSa.jpg)

[ 2 ] Protection Set

(IMG:[imgur.com] https://imgur.com/hLyM0Cu.jpg)

[ 3 ] Warding Set

(IMG:[imgur.com] https://imgur.com/rpLQqZM.jpg)

[ 4 ] Slaughter Set

(IMG:[imgur.com] https://imgur.com/bYNc8kd.jpg)

[ 5 ] Shielding Set

(IMG:[imgur.com] https://imgur.com/AS5pObX.jpg)

Here, as you can see, I was too lazy to spend a whole week with this set. (IMG:[invalid] style_emoticons/default/heh.gif) It's too slow. (IMG:[invalid] style_emoticons/default/faint.gif) But even so, you can already have a good idea of how this set behaves. (IMG:[invalid] style_emoticons/default/sleep.gif)

Simple conclusions:

The results of the Balance, Protection and Warding sets were very similar. And even considering that the Protection set is composed of 5 Peerless, and therefore presents the maximum numbers, the Balance set manages to be a little better due to the higher crit chance value, and the Warding set, as it consists of 5 Savage-type equipment (high crit damage), manages to get very close to the Protection set even with lower statistics.

In offensive terms, it is a little difficult to compare the Slaughter set with the others, due to the difference in the upgrade level (which is equivalent to something around 1300 ADB more if my set was forge level 25). But you can still see that the amount of ADB makes a lot of difference in the results (high damage/turn).

In defensive terms, of course nothing beats the Shielding set (a real walking fortress (IMG:[invalid] style_emoticons/default/heh.gif) ). But unfortunately it is too slow. (IMG:[invalid] style_emoticons/default/sad.gif) The Balance and Slaughter sets have less defenses, but in the case of Slaughter this is compensated by the greater attack power (that is, you kill the enemy faster and thus suffer less attacks). But regardless, as my overall numbers are high, in none of the tests did I need to use Cure or Health Potions. (IMG:[invalid] style_emoticons/default/happy.gif)

If someone is having defensive problems, I suggest mixing equipment types (Power only), and substituting for Slaughter as your defenses improve. (IMG:[invalid] style_emoticons/default/wink.gif)

In short, in terms of performance we have: Savage Slaughter > Slaughter >>> Savage Balance > Balance ≈ Savage Protection/Warding > Protection/Warding >>>>>> Shielding.

Tomorrow I'll post part 2 (if laziness not dominate me again (IMG:[invalid] style_emoticons/default/heh.gif) (IMG:[invalid] style_emoticons/default/heh.gif) (IMG:[invalid] style_emoticons/default/heh.gif) ).

This post has been edited by Basara Nekki: Nov 21 2020, 01:52

I knew about that, just didn't mention it. I think the bazaar price simply goes up linearly with the current condition (don't think if it's linear with maximum condition though).

I suppose so, but never bothered to check. If you want it's easy to check anyway.

How much exact condition does your shortsword have right now? Is it 290/377 or 291/377?

You're mixing up things. The condition is the number on the left in the weapon popup box, expressing how much it's damged and needs repairing; for my sword now is 611/800 (77%). I.e., the same number expressed in two formats, absolute and percentage. It will become 800/800 (100%) when repaired.

The xxx/377 you say is on the right, it is the Potency (PXP), and starts at 0 for unforged and un-IWed items ( 0/377 regardless of condition ). Each forging action or IW run will add some PXP points to it.

I know you know all this, but your post is a bit confused, speaking of condition but using numbers from Potency, which are unrelated.

My guess is that after you repair it, the bazaar sell price will be around 5404 credits (don't know which way rounding will go) but we'll see I hope. (IMG:[invalid] style_emoticons/default/smile.gif)

Well, I've rrepaired it now to check it (I would have done it anyway later) and it is 5420.

By the way, I forgot that soulfused/lottery equipment can't be sold to the bazaar (not sure) so I guess we can't check the sell prices using those.

Wrong, all item can be sold to bazaar, doesn't matter if they are soulbound or untradeable. They just cannot be sold to other player.

I should have posted these results before, but the laziness was stronger. (IMG:[invalid] style_emoticons/default/heh.gif)

As I did a lot of tests, I will divide it into three parts.

Part 1: Comparison between different "1H heavy sets".

-- skip --

Interesting, but ... for some reason, the imgur links don't work in my browser and I cannot see the images (IMG:[invalid] style_emoticons/default/sad.gif)

Anyway, the conclusions are not a great revolution; it is more or less what we could expect I'd say.

I'm waiting eagerly for part 2, with an humble prayer; may you use another site to link the images please ? (IMG:[invalid] style_emoticons/default/unsure.gif)

This post has been edited by mundomuñeca: Nov 18 2020, 03:34

Interesting, but ... for some reason, the imgur links don't work in my browser and I cannot see the images (IMG:[invalid] style_emoticons/default/sad.gif)

Anyway, the conclusions are not a great revolution; it is more or less what we could expect I'd say.

I'm waiting eagerly for part 2, with an humble prayer; may you use another site to link the images please ? (IMG:[invalid] style_emoticons/default/unsure.gif)

Weird. (IMG:[invalid] style_emoticons/default/huh.gif) For me Imgur has always worked. I already had problems with other image sites.

Which browser do you use? I use Iron, and I also have Firefox. The images appear correctly on both. Did you try opening the links directly in another browser tab?

Regarding the conclusions, they were purposely obvious. (IMG:[invalid] style_emoticons/default/tongue.gif) In fact, I wanted to see in numbers if there would be any significant differences between the sets. In addition, the numbers show that there is not much difference between a Protection and a Warding, when it comes to the defensive part. And yet, over time, most players have always preferred to buy Protection (most valued) over Warding (cheaper). Well, that could serve as a reference for new players, who are still putting together their sets.

I'm guessing Balance would do quite a bit better in DwD without Imperil, and even more so if using Vital Strike because it will let you Vital Strike sooner. It's no surprise that Balance would be unimpressive if anything other than non-imperil rapier is used for comparisons.

my sword now is 611/800 (77%)

Well, I've repaired it now to check it (I would have done it anyway later) and it is 5420.

Oops yeah I got mixed up there, thanks for correcting me. Strange that 611/800 = 76.375% rounds up to 77% but I guess that is how it goes. Actually there are multiple other weirdness regarding condition as well. I think fractional condition exists, unseen and rounded, because some of my equipment always drops 1 condition sooner than others.

Anyway 4161 * 800/611 = 5448.12 which is hmm, still not quite close enough to 5420 for me...

Let's try 4161 * 800/612 = 5439.22 yes I am disturbed haha. Hopefully I can figure this out.

Let's try 4160 * 800/612 = 5437.91 this is still no good, normal rounding can't perfectly explain this.

Whereas 4161 * 100/77 = 5403.90 isn't exactly right either...need to wait until I think of an idea. Seems roughly linear at least.

I'm guessing Balance would do quite a bit better in DwD without Imperil, and even more so if using Vital Strike because it will let you Vital Strike sooner. It's no surprise that Balance would be unimpressive if anything other than non-imperil rapier is used for comparisons.
Oops yeah I got mixed up there, thanks for correcting me. Strange that 611/800 = 76.375% rounds up to 77% but I guess that is how it goes. Actually there are multiple other weirdness regarding condition as well. I think fractional condition exists, unseen and rounded, because some of my equipment always drops 1 condition sooner than others.

Anyway 4161 * 800/611 = 5448.12 which is hmm, still not quite close enough to 5420 for me...

Let's try 4161 * 800/612 = 5439.22 yes I am disturbed haha. Hopefully I can figure this out.

Let's try 4160 * 800/612 = 5437.91 this is still no good, normal rounding can't perfectly explain this.

Whereas 4161 * 100/77 = 5403.90 isn't exactly right either...need to wait until I think of an idea. Seems roughly linear at least.

The percentage value is to highest integer to make sure it reaches 100% and for display reason (field length), the internal value is fractional for sure since it is calculated from the two values 611/800 like you did.
A completely spent item would have 1 / 800 (1%) values. Zero values are most probably banned to avoid strange exceptions effect in a convoluted informatical environment (IMG:[invalid] style_emoticons/default/laugh.gif) (IMG:[invalid] style_emoticons/default/tongue.gif)

As for the function, even if the slope is linear, you forget that it's not a zero-crossing function. There is still a residual sell value when the condition is zero (actually, lowest possible condition is 1%).

But I'm not going to make my weapon reach 1% condition to tell you how much that residual value is, sorry (IMG:[invalid] style_emoticons/default/tongue.gif)

EDIT : just for example, and using completely made-up fake numbers to make it simple, suppose that the residual value @ 1% condition is 1460 C.

Then the value increment from 1% to 100% would be (5420 - 1460) = 3960 Now each 1% increment in condition would give (3960 / 99) = 40 C

Under this hyphotesis a 77% condition would give 1460 + (77 * 40) = 4540 C

You can do it in reverse (find the equation parameters from two point it's elementary), but for better accuracy it would be better to use a difference as high as possible between the values, i.e. the value @ 100% condition and the value @ 1% condition for the same item (using as second point the value @ 77% condition would give higher errors in the parameters estimation).

This post has been edited by mundomuñeca: Nov 18 2020, 15:19